This review emphasizes the predictive ability, sensitivity and specificity of aquatic plant biomarkers as biomonitoring agents of exposure and effect. Biomarkers of exposure are those that provide functional measures of exposure that are characterized at a sub-organism level. Biomarkers of effect require causal linkages between the biomarker and effects, measured at higher levels of biological organization. With the exception of pathway specific metabolites, the biomarkers assessed in this review show variable sensitivity and predictive ability that is often confounded by variations in growth conditions, rendering them unsuitable as stand alone indicators of environmental stress. The use of gene expression for detecting pollution has been, and remains immature; this immaturity derives from inadequate knowledge on predictive ability, sensitivity and specificity. Moreover, the ability to the detect mode of action of unknown toxicants using gene expression is not as clear-cut as initially hypothesized. The principal patterns in gene expression is not as clear-cut as initially hypothesized. The principal patterns in gene expression are generally derived from stress induced genes, rather than on ones that respond to substances with known modes of action (Baerson et al. 2005). Future developments in multivariate statistics and chemometric methods that enhance pattern analyses in ways that could produce a "fingerprint", may improve methods for discovering modes of action of unknown toxicants. Pathway specific metabolites are unambiguous, sensitive, correlate well to growth effects, and are relatively unaffected by growth conditions. These traits make them excellent biomarkers under both field and laboratory conditions. Changes in metabolites precede visible growth effects; therefore, measuring changes in metabolite concentrations (Harring et al. 1998; Shaner et al. 2005). The metabolic phase I enzymes (primarily associated with P-450 activity) are non-specific biomarkers, and few studies relate them to growth parameters. P-450 activity both increases and decreases in response to chemical stress, often confounding interpretation of experimental results. Alternatively, phase II metabolic enzymes (e.g., glutathione S-transferases; GST's) appear to be sensitive biomarkers of exposure, and potentially effect. Some GST's are affected by growth factors, but others may only be induced by xenobiotics. Measuring xenobiotic-induced GST's, or their gene expression patterns, are good candidates for future biomarkers of the cumulative load of chemical stress, both in the laboratory and under field conditions. Phytochelatins respond to some but not all metal ions, and may therefore be used as biomarkers of exposure to identify the presence and bioavailability of ions to which they respond. However, more data on their specificity to, and interactions with growth factors, in more species are needed. The flavenoids are only represented by one heavy metal exposure study; therefore their use as biomarkers is currently difficult to judge. S...